Financial Applications ofClassification and Regression Trees

Abstract

This study gives an outline of modern theory of classification and regression trees (CART) and shows the advantages of CART applications in finance. Practical issues regarding CART applications and core implementation are presented. The second part of the work is mainly concentrated on DAX30 market simulation results and shows how a CART-based business application can perform on stock market as well as what supplementary results can be got using CART as a forecasting system. In this realm comparison of technical and fundamental approaches is performed. Finally, information ageing effect in the context of learning sample construction is analyzed.